Savita Biradar

Project: IDG summary page powered by GenAI for hospice care — managed data migration, ETL development, and implemented AI-driven summaries while handling sensitive patient PII.

• Migrated data from SQL Server to PostgreSQL in Azure, ensuring data integrity and consistency.
• Developed and deployed ETL processes in Azure Function Apps with time-based triggers.
• Preprocessed staging data before integrating into the IDG schema to improve data readiness.
• Conducted data modeling to optimize database performance and scalability.
• Built the IDG summary page using OpenAI GPT-4o for AI-generated clinical summaries.
• Deployed models using Azure DevOps pipelines to automate and standardize deployments.
• Added edit/save functionality to capture user feedback for reinforcement learning and fine-tuning.

Environment: GenAI, Azure Function Apps, Azure Event Hub, PostgreSQL, SQL Server, ETL, Data Modeling, OpenAI GPT-4o, Azure DevOps

Led platform decisions and managed Dataiku DSS environments, and designed NLP and generative AI solutions across multiple projects.

• Managed Dataiku DSS setup, maintenance, upgrades, disk and memory tuning, and user permissions.
• Coordinated releases, backups, and migration of DSS instances; automated workflows and code environments.
• Built scalable NLP pipelines, training workflows, and low-latency inference engines (PyTorch/TensorFlow).
• Implemented model fine-tuning, distillation, pruning, and CUDA optimizations to reduce inference latency.
• Automated wealth-management ETL flows using Dataiku and optimized SQL for production readiness.
• Developed ESG data pipelines on AWS using PySpark (EMR), Glue, Lambda, and Redshift.

Environment: Linux, Terminal, Impala, Spark SQL, Python, Dataiku, HiveQL, PyTorch, TensorFlow, TensorFlow Serving, ONNX Runtime, AWS, EMR, S3, Glue

Worked on cloud migrations and data lake transformations, including dbt orchestration and Hadoop-to-Azure data movement.

• Configured dbt to connect source (IBM Cloud) and target (Azure Data Lake) and developed dbt models.
• Designed data transformations to meet business rules and implemented migration workflows.
• Analyzed large data sets using Hive, created tables on Hadoop data lakes, and wrote MapReduce/Hive jobs.

Environment: dbt, SQL, Azure Data Lake, Hadoop, HDFS, Hive, PySpark, AWS, EMR, S3, Lambda

Developed a dialogue system to communicate breast cancer information to Latinx patients using NLP.

• Implemented NLP experiments using BERT and cosine similarity models and compared results to human judgments.
• Built backend services with Flask and used TensorFlow and Transformers for model development.
• Used SNLI corpus for semantic evaluation and applied Python, NumPy, and ML toolchains.

Environment: Python, TensorFlow, Flask, ML, NLP, Anaconda, R

Integrated AI into talent acquisition systems and built ML/NLP models and APIs.

• Developed ML models (LDA, Naive Bayes, PCA) and applied data cleaning, quality checks, and feature engineering.
• Built REST APIs, API tests, and automation scripts using Python and Selenium.
• Participated in product meetings and provided performance feedback and improvement suggestions.

Environment: Python, scikit-learn, Gensim, Data Studio, NLP, MongoDB, Kafka

Enhanced search relevance using NLP, trained SpaCy NER models, and deployed via Flask APIs.

• Created training data for NER tagging and trained SpaCy models; deployed models with Flask APIs.
• Built Python web crawlers with Scrapy and created NIFI workflows to run them; pushed processed data to Elasticsearch.
• Built dashboards and reports in Google Data Studio to demonstrate results and quick wins.

Environment: Python, SpaCy, NLTK, Scrapy, Apache NIFI, Google Data Studio